目的：系统性阐明解脂耶氏酵母（Yarrowia lipolytica）侵袭性感染的临床特征及实验室鉴定情况；建立高分辨率的分子分型方法，结合全基因组数据分析开展分子流行病学调查；综合评价体外抗真菌药物敏感性特征并建立流行病学界值，同时探究解脂耶氏酵母对唑类药物耐药的分子机制；尝试构建动物模型并评价其毒力因子及致病机制，为临床诊疗及院内感染防控提供科学依据。

方法：收集中国医院侵袭性真菌病监测网（China Hospital Invasive Fungal Surveillance Net, CHIF-NET）2009年至2019年间来自21家监测中心的55株解脂耶氏酵母临床分离株，回顾性分析病例及感染特征，同时收集22株非临床来源菌株作为参考。实验室鉴定采用Sanger测序和基质辅助激光解吸电离飞行时间质谱（Matrix-Assisted Laser Desorption Ionization Time-Of-Flight Mass Spectrometry, MALDI-TOF MS）技术，对各监测分中心鉴定结果及误判情况进行综合评估。在分子流行病学研究方面，利用全基因组单核苷酸多态性（Single Nucleotide Polymorphism, SNP）分析、核糖体DNA序列分析、建立多位点序列分型（Multilocus Sequence Typing, MLST）和简单重复序列（Simple Sequence Repeat, SSR）分型方法，进行遗传流行病学特征系统分析。采用BMD、Sensititre YeastOne™、ATB^® FUNGUS 3以及 Liofilchem最低抑菌浓度试纸条（MTS）检测9种抗真菌药物的敏感性，进而建立本地流行病学界值（Local Epidemiological Cut-Off Value, L-ECOFF）。利用流式细胞术、全基因组测序和转录组测序等技术深入探讨唑类耐药机制。致病机制方面，通过压力试验、体外毒力因子检测、免疫功能正常及受抑制小鼠感染模型评估菌株毒力，检测感染小鼠各器官真菌负荷、生存曲线，细胞因子及组织病理学改变；基于同源臂介导双交换同源重组，验证解脂耶氏酵母分泌型天冬氨酸蛋白酶（Secreted Aspartic Protease, SAP）的调控基因，并评价其对毒力和耐药性的影响。

结果：流行病学调查显示，解脂耶氏酵母感染患者均无免疫缺陷病史但多有手术外伤史、长期住院和侵入性操作史。患者平均年龄为53.1岁，男性发病率显著高于女性（2.9:1）。血流感染是最主要的感染类型（65.4%），其中导管相关血流感染占20.6%。约48.1%的患者表现出呼吸系统感染症状，且在血流感染患者中，55.9% 同时伴有下呼吸道感染或重症肺炎。据统计，解脂耶氏酵母感染的全因死亡率达 21.2%。实验室鉴定结果表明，分中心鉴定的准确率为72.73%。传统鉴定方法的存在误判情况，MALDI-TOF MS鉴定解脂耶氏酵母中的准确率为100%。基于全基因组SNP分析，揭示临床致病菌株与非临床分离菌株在基因组层面无显著差异；在国内建立并优化解脂耶氏酵母的MLST方法，修正了国际研究中存在的序列错误及引物适配性问题；开发解脂耶氏酵母的SSR标记分型技术，筛选并确定了4个特异性强、敏感度高的SSR分子标记。基于SSR快速筛查、MLST进化分型和SNP高分辨率鉴定三级分层分析，发现中国H01医院重症监护室分离菌株聚集为两个明显的单克隆传播簇，簇内SNP差异仅为421–458 bp，而簇间的SNP差异高达45,315–45,389 bp。抗真菌药物敏感性研究发现，解脂耶氏酵母对棘白菌素类药物普遍敏感，BMD法测得的L-ECOFFs为1–2 µg/mL，均为野生型。但唑类药物的非野生型分离株占比16.22–18.92%，其中氟康唑L-ECOFFs为8 µg/mL，非野生型占比18.92%。分析不同药敏方法在检测临床分离菌株间的基本一致性发现，ATB^®法与BMD法的基本一致性（83.64–100%）最高，其次YeastOne™（61.82–100%）和MTS（47.27–98.18%）。耐药机制研究表明，唑类药物非野生型菌株在转录组水平上呈现显著的表达模式变化，主成分分析显示其与野生型菌株的空间分布明显分离。氟康唑非野生型临床分离株携带YALI0_B05126g基因A395T纯合突变，而非临床分离株未检测到该突变。基因组结构分析显示，非野生型菌株的染色体稳定，无大规模染色体重排或非整倍体现象，但YALI0_B05126g基因的拷贝数发生变异，且变异情况与唑类药物MIC 值呈中等至强正相关，尤其是与泊沙康唑MIC值的相关性最高（rho = 0.87）。此外，生物膜的形成可显著提高解脂耶氏酵母对唑类药物的MIC值，且非野生型分离株的生物膜形成能力更强（P < 0.05）。毒力机制研究发现，解脂耶氏酵母可分泌多种胞外酶，其中SAP是其重要的毒力因子，且哺乳动物体温下SAP活性相比环境温度显著增强（P < 0.05）。YALI0_A16819g基因对SAP的表达具有负向调控作用，并可介导解脂耶氏酵母由酵母型向假菌丝或菌丝形态的转换。成功构建的免疫功能正常及免疫抑制小鼠血流感染模型进一步证实，解脂耶氏酵母主要侵犯肺脏，其次为脾脏、肾脏、肝脏和脑组织，肺部真菌负荷可作为菌株毒力的重要评价指标。血清可诱导解脂耶氏酵母的菌丝发育，增强其体内存活能力。此外，长期或强烈的唑类药物压力可诱导SAP酶活性上调，以增强解脂耶氏酵母对药物环境的适应性和生存能力。

结论：本研究回顾性调查解脂耶氏酵母感染的流行病学特征，构建了基于SSR快速筛查、MLST进化分型和SNP高分辨率鉴定的三级院感监测预警模型，为早期发现和控制解脂耶氏酵母的传播提供了重要分子工具。首次建立解脂耶氏酵母的L-ECOFFs，揭示解脂耶氏酵母唑类耐药的关键机制，包括YALI0_B05126g基因突变、拷贝数变异及生物膜形成，为临床合理用药提供了依据。通过构建小鼠血流感染模型，明确该病原的关键毒力因子和SAP毒力相关机制，发现YALI0_A16819g基因对于解脂耶氏酵母菌丝发育及毒力增强中的关键作用，为后续防控与靶向治疗奠定基础。

Objective: To systematically elucidate the clinical characteristics and laboratory identification of invasive infections caused by Yarrowia lipolytica; to establish a high-resolution molecular typing method and conduct a molecular epidemiological investigation using whole-genome data analysis; to comprehensively assess the in vitro antifungal susceptibility profile, establish epidemiological cutoff values, and explore the molecular mechanisms underlying azole resistance in this pathogen; and to attempt the construction of an animal model to evaluate its virulence factors and pathogenic mechanisms, thereby providing scientific evidence for clinical diagnosis, treatment, and hospital infection control.

Methods: Clinical isolates of Y. lipolytica were collected from 21 surveillance centers participating in the China Hospital Invasive Fungal Surveillance Net (CHIF-NET) between 2009 and 2019. A retrospective analysis of clinical cases and infection characteristics was conducted, with additional non-clinical isolates included as references. Laboratory identification was performed using Sanger sequencing and Matrix-Assisted Laser Desorption Ionization Time-Of-Flight Mass Spectrometry (MALDI-TOF MS), and the accuracy of species identification across surveillance centers was comprehensively evaluated. For molecular epidemiology studies, genetic epidemiological characteristics were systematically analyzed using whole-genome single nucleotide polymorphism (SNP) analysis, ribosomal DNA sequencing, and the establishment of multilocus sequence typing (MLST) and simple sequence repeat (SSR) typing methods. Antifungal susceptibility testing for nine antifungal agents was conducted using broth microdilution (BMD), Sensititre YeastOne™, ATB^® FUNGUS 3, and MIC test strip (MTS) methods, leading to the establishment of local epidemiological cutoff values (L-ECOFFs). Azole resistance mechanisms were explored using flow cytometry, whole-genome sequencing, and transcriptome sequencing. Pathogenicity studies included in vitro virulence factor assays and the evaluation of virulence in both immunocompetent and immunosuppressed mouse infection models. Fungal burden in various organs, survival curves, cytokine responses, and histopathological changes in infected mice were assessed. Furthermore, homologous arm-mediated double-crossover homologous recombination was employed to validate the regulatory genes of Y. lipolytica secreted aspartic protease (SAP), and its impact on virulence and antifungal resistance was evaluated.

Results: The epidemiological investigation revealed that all patients with Y. lipolytica infections had no history of immunodeficiency but commonly had a history of surgical trauma, prolonged hospitalization, and invasive procedures. The mean patient age was 53.1 years, with a significantly higher incidence in males than females (2.9:1). Bloodstream infection (BSI) was the most common infection type (65.4%), with catheter-related BSIs accounting for 20.6%. Approximately 48.1% of patients exhibited respiratory infection symptoms, and among those with BSI, 55.9% also had lower respiratory tract infections or severe pneumonia. The all-cause mortality rate of critically ill patients with Y. lipolytica infection was 21.2%. Laboratory identification results showed that the accuracy of species identification at surveillance centers was 72.73%. Traditional identification methods had misidentifications, whereas MALDI-TOF MS achieved 100% accuracy in identifying Y. lipolytica. In contrast, MALDI-TOF MS achieved 100% accuracy in identifying Y. lipolytica. Whole-genome SNP analysis revealed no significant genomic differences between clinical and non-clinical isolates. The study established and optimized an MLST method for Y. lipolytica in China, correcting sequence errors and primer mismatches present in international studies. Additionally, an SSR-based typing method was developed, identifying four highly specific and sensitive SSR markers. A three-tiered hierarchical approach—rapid SSR screening, MLST-based evolutionary typing, and high-resolution SNP analysis—revealed two distinct monoclonal transmission clusters of Y. lipolytica in the ICU of Hospital H01 in China. The SNP differences within clusters ranged from 421 to 458 bp, while inter-cluster differences reached 45,315-45,389 bp. Antifungal susceptibility testing demonstrated that Y. lipolytica was universally susceptible to echinocandins, with L-ECOFFs of 1–2 µg/mL determined by the BMD method, all within the wild-type. However, the proportion of non-wild-type isolates for azole drugs ranged from 16.22% to 18.92%, with fluconazole L-ECOFFs set at 8 µg/mL and a non-wild-type rate of 18.92%. Comparative analysis of susceptibility testing methods in clinical isolates showed the highest essential agreement between BMD and ATB^® (83.64–100%), followed by YeastOne™ (61.82–100%) and MTS (47.27–98.18%). Studies on azole resistance mechanisms revealed significant transcriptomic expression changes in non-wild-type isolates, with principal component analysis demonstrating clear spatial separation between wild-type and non-wild-type strains. Fluconazole non-wild-type clinical isolates harbored a homozygous A395T mutation in the YALI0_B05126g gene, which was absent in non-clinical isolates. Genomic structural analysis showed chromosomal stability in non-wild-type strains, with no large-scale chromosomal rearrangements or aneuploidy. However, copy number variations of YALI0_B05126g correlated moderately to strongly with azole MIC values, with the highest correlation observed for posaconazole (rho = 0.87). Additionally, biofilm formation significantly increased azole MIC values, and non-wild-type isolates exhibited stronger biofilm-forming ability (P < 0.05). Virulence mechanism studies revealed that Y. lipolytica secretes multiple extracellular enzymes, with SAP identified as a key virulence factor. Notably, SAP activity was significantly enhanced at mammalian body temperature compared to environmental temperature (P < 0.05). The YALI0_A16819g gene negatively regulated SAP expression and mediated yeast-to-hyphal morphological transitions. Immunocompetent and immunosuppressed mouse bloodstream infection models confirmed that Y. lipolytica primarily infects the lungs, followed by the spleen, kidneys, liver, and brain, with pulmonary fungal burden serving as a key virulence indicator. Serum exposure induced hyphal development, enhancing in vivo survival. Furthermore, prolonged or intense azole pressure upregulated SAP activity, facilitating adaptation and survival in drug-exposed environments.

Conclusion: This study retrospectively investigated the epidemiological characteristics of Y. lipolytica infections and developed a three-tier hospital infection surveillance and early warning model based on SSR rapid screening, MLST evolutionary typing, and high-resolution SNP identification. This model provides essential molecular tools for the early detection and control of Y. lipolytica transmission. For the first time, L-ECOFFs for Y. lipolytica were established, uncovering key mechanisms of azole resistance, including YALI0_B05126g gene mutations, copy number variations, and biofilm formation, providing a basis for rational antifungal therapy. By constructing a murine bloodstream infection model, this study identified critical virulence factors and SAP-related virulence mechanisms. The YALI0_A16819g gene was found to play a key role in hyphal development and virulence enhancement, laying the foundation for future prevention strategies and targeted therapies.